a multi‐physical simulation architecture to support the regenerative braking. • competing...

IPG Technology ConferenceKarlsruhe 2012

A multi physical simulation architecture to support theA multi‐physical simulation architecture to support the development of hybrid electric vehicles

James Chapman

CAE Simulation GroupJaguar Land Roverg

Embedded Systems GroupWMG – University of Warwick

LCVTP Systems Model Framework

ContentsContents

• IntroductionIntroduction

> Objectives, Scope & Use Cases

• Simulation Platform RequirementsSimulation Platform Requirements

• Model Structure

> C t l/Pl t M d l A hit t> Control/Plant Model Architecture

> IPG CarMaker Integration

> HIL (Hardware in the Loop) Environment( p)

• Example Application: Regenerative Braking & Vehicle Dynamics

• ConclusionsConclusions


IntroductionIntroduction

LCVTP: Multi‐partner collaborative project with an aim to progress the development of low carbon vehicle technology within the West Midlandsdevelopment of low carbon vehicle technology within the West Midlands.

Systems Modelling Activities:

• Vehicle systems simulation platform developed to support integrated model‐based development activities across a range of different project workstreams.

• Objective to develop a single common simulation environment to promote collaborative development & dissemination of models among multiple project partners.

• Intended to accommodate a number of different use cases through the b f b h l l b hsubstitution of both plant & control subsystem components with varying

degrees of functionality and fidelity.


Project ScopeProject Scope

Example Use Cases:

• Model based development of control systems for low carbon vehicles.

> VSC – Vehicle Supervisory Control

> HVAC & Cooling Advanced Thermal Energy Management Systems> HVAC & Cooling – Advanced Thermal Energy Management Systems

> BMS – Battery Management System (High Voltage)

> Regenerative Braking & Stability Control

• Study the impact of regenerative braking on vehicle dynamics.

• Performance & fuel economy predictions for competing HEV architectures.Performance & fuel economy predictions for competing HEV architectures.

> Incl. potential benefit from thermal energy recovery systems & minimisation of parasitic losses.

• Component sizing: e.g. drive motor, APU & energy storage units.

• Optimisation of powertrain cooling systems.


Platform RequirementsPlatform Requirements

• Support a wide range of emerging hybrid electric vehicle architectures.

> Requires capability for multi‐physical modelling of powertrain.

• Development of high/low voltage electrical networks.

> Accommodate a range of different electrical architectures, subsystems & analysis.

> Requires consideration of electrical dependency in model derivation & physical> Requires consideration of electrical dependency in model derivation & physical connections between HV/LV electrical subsystems.

• Development of thermal network.

> Inclusion of thermal management system.

> Requires inclusion of thermal dependency in subsystem models.

> Support multiple interactions between thermally dependent subsystems.


Platform RequirementsPlatform Requirements

• Required to support MATLAB/Simulink derived controllers• Required to support MATLAB/Simulink derived controllers.

• To include an appropriate realisation of the full vehicle model for real time

simulation within a HIL (Hardware In Loop) environmentsimulation within a HIL (Hardware‐In‐Loop) environment.

• Framework required to extend into existing IPG CarMaker model environment to

study vehicle dynamicsstudy vehicle dynamics.

• Vehicle level models to be derived from standard libraries of subsystem models to

promote parallel development & maximise reuse of core assetspromote parallel development & maximise reuse of core assets.

• Appropriate SVN version control & standardised signal naming convention for

project wide collaborationproject‐wide collaboration.


Model StructureModel Structure


Model Structure

Framework includes three top‐level implementations constructed from a

Model Structure

single library of subsystem components.

1. Simulink Longitudinal

• Simulink based control models + embedded Dymola powertrain model.

• Example use cases – fuel efficiency studies, controller development etc.

2. Dymola Standalone

• Basic control functionality + drive cycle implemented in standalone Dymola model.

• Example use cases – HV electrical transients, vehicle dynamics etc.

3. CarMaker

• CarMaker extension of Simulink longitudinal inc. CarMaker suspension & tyre models.

• Example use cases – vehicle dynamics, HIL simulation + failsafe tester etc.


Model Structure – Control ArchitectureModel Structure – Control Architecture

C l hi b d JLR V hi l• Control architecture based on JLR Vehicle Model Architecture (VMA).

• Subsystems modularised according to workstream activities & vehicle ECUworkstream activities & vehicle ECU deployment.

• Dymola powertrain integrated using standard Simulink s‐function interface.

• Plant subsystems lumped to maintain acausalinteractions between componentsinteractions between components.


Model Structure – Plant ArchitectureModel Structure – Plant Architecture

• Dymola HEV vehicle model architecture.

> High & low voltage electrical networks. g g

> Thermal interactions between subsystems.

• Object orientated structure.


Model Structure – IPG CarMaker IntegrationModel Structure – IPG CarMaker Integration

Complete powertrain model including plant & control integrated into CarMaker vehicle dynamics model via Simulink interface using DVA (Direct Variable Access).


Model Str ct re HIL Real Time En ironment

• Real‐time simulation of LCVTP full vehicle model

Model Structure – HIL Real‐Time Environment

implemented on CarMaker XPack4 HIL Platform

• HIL platform used to prove capability of control algorithms & diagnostics to run in real‐time on 32bit processor platform.

• Also used to study the impact of y psignal propagation delays over CAN & Flexray networks, & robustness to fault injection.j


Model Structure – HIL Real Time EnvironmentModel Structure – HIL Real‐Time Environment

• Executed on IPG XPack4 real‐time computer.

• Specific controllers split out onto independent ECUs

• Complete powertrain model, including plant & control integrated into CarMaker HIL model independent ECUs. control, integrated into CarMaker HIL model.


Example Use Cases – Regenerative BrakingExample Use Cases – Regenerative Braking

Objective: To support the evaluation & development of integrated control systems associated with

advanced regenerative braking.

• Competing electro‐hydraulic braking systems.

• Impact of brake system configuration on potential energy recovery & overall braking performance.

• Control strategy (interaction with ABS).

• Immediate vs. blended termination of regenin response to stability event.

• Control architecture.

• Impact of specific ECU deployment & latency signal propagation delayslatency signal propagation delays.



Vehicle‐level model based testing performed using medium fidelity powertrain subsystem models & C k /Si li k i fCarMaker/Simulink interface.

• 1st order hydraulic brake models implemented in Dymola incl Electro‐implemented in Dymola incl. Electrohydraulic regen brakes & ABS modulation.

• Neglection of 2nd order characteristics.

• Incompressible hydraulic fluid.

• Lumped compliance for pipes, pads & callipers.

• Linearisation of discontinuities.

• Suitable for real time application.



Control Architecture Case Study:

• BTAC – Brake Torque Apportionment Controller responsible for arbitration of friction & e‐machine braking torques for a given driver demand.

• Design consideration: communication network & critical deployment of key functions onto vehicle ECUs.



Front axle regenerative braking with handover to friction braking at <35kph



Impact of signal propagation delays due to varying CAN message cycle times.



Standard foundation brake system response to high‐low grip event with & without ABS.



Impact of regenerative braking with varying control architecture & transit delays.


ConclusionsConclusions

• Hybrid‐electric vehicle systems are inherently multi‐physical & require a more integrated systems approach to model based development.

• A vehicle systems simulation platform has been developed as part of the y p p pLCVTP to support integrated systems modelling activities across a range of different technology areas.

• Through the substitution of analogous library models with varying levels of functionality & fidelity, an array of diverse use cases may be addressed.

• IPG CarMaker & XPack4 have used in conjunction with MATLAB/Simulink & Dymola physical modelling package for HEV control development.

• For more information on the project see: http://www2.warwick.ac.uk/fac/sci/wmg/research/lcvtp

a multi‐physical simulation architecture to support the regenerative braking. • competing...

Documents